AI agents for data cleaning
Deduplicate, normalise and enrich records before they reach a system of record.
6 agents, each with its inputs, outputs and review level checked by a person.
Data cleaning agents deduplicate, normalise and enrich records before they reach a system of record: contacts, companies, products, transactions. They take a file or a connection, apply rules and lookups, flag what they could not resolve and hand back a clean set with a log of what changed. The listings record the sources each agent connects to, the operations it performs, what it returns and whether a person reviews changes before they are written back.
Use them where bad data costs money downstream, such as CRM imports and product catalogues. Check whether the agent writes back automatically or produces a reviewed file, and how it handles ambiguity.
The agents
- Claygentby ClayRuns a research prompt against every row of a lead table and returns structured data.Freemium, from $167/moNeeds configurationOptional reviewDetails
- DataLabby DataCampAI data notebook that writes SQL and Python from a question and returns a shareable report.FreemiumNeeds configurationHuman reviewDetails
- Landbaseby LandbaseFinds, qualifies and ranks target accounts with agents, then enriches the contacts.Usage-based, from $499/moNeeds configurationOptional reviewDetails
- monday.com AI agentsby monday.comPrebuilt and custom agents that research, report and flag risks inside monday.com boards.Subscription, from €9 per seat/moNeeds configurationOptional reviewDetails
- Rows AI Analystby RowsSpreadsheet AI that writes formulas, builds charts and analyses data from a typed prompt.Freemium, from $8 per user/moWorks instantlyHuman reviewDetails
- Vizlyby VizlyUploads a data file, answers questions in plain language and writes the Python or R behind each chart.FreemiumWorks instantlyHuman reviewDetails
No agent matches. Clear the search or choose another filter.
Related
Groups that share agents with this one.
Questions
Will the agent overwrite my data?
Only if configured to write back. Many produce a cleaned file or a change log for review; the review level on each listing reflects the vendor's default.
Where does enrichment data come from?
From data providers or the web, as each listing's integrations and inputs describe. The catalogue does not verify coverage claims.